Thank you for adding the AI terms link.

Per our submission guidelines, this page must remain flagged for now. Neural.love talks about practicing "Ethical AI training", but these practices are about unbiased models, not about ethically sourced training data. Neural.love's training data still comes from huge undocumented corpuses that most likely scraped images with uncertain copyrights. Because we don't yet know how these datasets will be treated by emerging intellectual property law, we cannot yet trust their claim that outputs are all CC0. We're hoping that's true, but it may be like saying "Our quality knives are made from other people's stolen knives that have since been melted down and reforged. Because of this, we don't claim any ownership over them and we're giving them away for free!"

See also https://opengameart.org/content/artificial-intelligence-assisted-artwork